Applied Mechanics and Materials Vols. 427-429

Abstract: Particle swarm optimization meets difficulties when handling deceptive problems, because the local optimal attractors misguide the particles and pull them away from the global optimum. Function transform that changes the shape to make it easier for the algorithm to optimize is an efficient way to change the original attraction, but the existing strategies cannot conquer deception. Therefore we propose the mirror transform by simply reversing the original attraction. Experimental results validate the efficiency of our strategy.

Abstract: In this paper, we propose a novel method for image retrieval based on multi-instance learning with relevance feedback. The process of this method mainly includes the following three steps: First, it segments each image into a number of regions, treats images and regions as bags and instances respectively. Second, it constructs an objective function of multi-instance learning with the query images, which is used to rank the images from a large digital repository according to the distance values between the nearest region vector of each image and the maximum of the objective function. Third, based on the users relevance feedback, several rounds may be needed to refine the output images and their ranks. Finally, a satisfying set of images will be returned to users. Experimental results on COREL image data sets have demonstrated the effectiveness of the proposed approach.

Abstract: Image registration is widely used in applications for mapping one image to another. As it is often formulated as a point matching problem, in this paper, a novel method, called the Geometric Inference (GI) algorithm, is proposed for feature point based image registration. Firstly, according to affine distance invariant, the global geometric relationship between collinear correspondences is deduced and used for collinear point matching. Secondly, utilizing affine area invariant, geometric relationship between noncollinear correspondences is inferred and used for noncollinear point matching. Finally, the best affine transformation can be discovered from the correspondences composed of the collinear and noncollinear corresponding point pairs. Experiments on synthesized and real data demonstrate that GI is well-adapt to image registration as it is fast and robust to missing points, outliers, and noise.

Abstract: The traditional structure learning algorithms are mainly faced with a large sample dataset. But the sample dataset practically is small. Based on it, we introduce the Probability Density Kernel Estimation (PDKE), which would achieve the expansion of the original sample sets. Then, the K2 algorithm is used to learn the Bayesian network structure. By optimizing the kernel function and window width, PDKE achieves the effective expansion of the original dataset. After the confirm of variable order based on mutual information, a small sample set of Bayesian structure learning algorithm would be established. Finally, simulation results confirm that the new algorithm is effective and practical.

Abstract: The present study is concerned about image mosaic in single reflector panoramic imaging system (SRPIS). A nonlinear image mosaic algorithm is proposed to get the panoramic image of pipe inner surface. Because of nonlinear distortion in the images which are unwrapped from the original images, its practically impossible for traditional image mosaic method based on 2D planar projective transformation to eliminate phenomenon of ghost and blur in the seam. Nonlinear image mosaic algorithm is performed by projecting many pieces of image divided from right image onto the left image. The position-variant parameters of transformation model are got by quadratic interpolation. The results show that nonlinear image mosaic algorithm overcomes the limitations of traditional image mosaic method in images with distortion and the mosaic image is clearer than that by traditional image mosaic method.

Abstract: The present study concerns about feature matching in image mosaic. In order to solve the problems of low accuracy and poor applicability in the traditional speeded up robust features algorithm, this paper presents an improved algorithm. Clustering algorithm based on density instead of random sample consensus method is used to eliminate mismatching pairs. The initial matching pairs are mapped onto a plane coordinate system, which can be regarded as points, by calculating the density of each point to extract the final matching pairs. The results show that this algorithm overcomes the limitations of the traditional speeded up robust features mosaic method, improving the matching accuracy and speed, and the mosaic effect. It has certain theoretical and practical value.

Abstract: Classification on geospatial data is different from classical classification in that spatial context must be taken into account. In particular, the validation criterion functions should incorporate both classification accuracy and spatial accuracy. However, direct combination of the two accuracies is cumbersome, due to their different subjects and scales. To circumvent this difficulty, we develop two new criterion functions that indirectly incorporate spatial accuracy into classification accuracy-based functions. Next, we formally introduce a set of ideal properties that an appropriate criterion function should satisfy, giving a more meaningful interpretation to the relative significance coefficient in the weighted scheme. Finally, we compare the proposed new criterion functions with existing ones on a large data set for 1980 US presidential election.

Abstract: Real-time plane reflection is shown by adding a mirror camera. Roaming scene, capture the mirror image of the scene and the mirror image map to the reflection plane by projection. The multi-layer superimposed image is processed on the reflection plane, by different of superposition algorithms, different results obtained, that displays the effect of different plane reflection.

Abstract: A new improved image fusion algorithm is proposed for multi-spectral image (MUL) and the high-resolution panchromatic image (PAN) based on intensity-hue-saturation (IHS) transform combined with wavelet transformation (WT). Firstly, the multi-spectral image is transformed into the IHS space for getting the intensity component (I).Then the high-resolution panchromatic image and I were matched with histogram. Secondly, the PAN and I were decomposed respectively by WT and fused to obtain the new I by the inverse WT. Finally, the fusion image was obtained by inverse IHS transform. four evaluate indicators are defined in this paper. By experiment research, the results show that this new method can effectively improve the fusion effect.

Abstract: Digital design has resulted in the need for a reconfigurable of current design theories. The present research postulates the oretical framework of informed tectonics in digital design. The research for the evolution algorithmgene expression programming is developed and discussed. This paper mainly talk about the oretical framework of informed tectonics in digital design based on evolution algorithmGEP. The methodologies is used to explain and guide future research anddevelopment.